[mirror_ubuntu-zesty-kernel.git] / fs / proc / inode.c

/*
 *  linux/fs/proc/inode.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

#include <linux/time.h>
#include <linux/proc_fs.h>
#include <linux/kernel.h>
#include <linux/pid_namespace.h>
#include <linux/mm.h>
#include <linux/string.h>
#include <linux/stat.h>
#include <linux/completion.h>
#include <linux/poll.h>
#include <linux/printk.h>
#include <linux/file.h>
#include <linux/limits.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/sysctl.h>
#include <linux/seq_file.h>
#include <linux/slab.h>
#include <linux/mount.h>

#include <asm/uaccess.h>

#include "internal.h"

static void proc_evict_inode(struct inode *inode)
{
	struct proc_dir_entry *de;
	struct ctl_table_header *head;
	const struct proc_ns_operations *ns_ops;
	void *ns;

	truncate_inode_pages(&inode->i_data, 0);
	clear_inode(inode);

	/* Stop tracking associated processes */
	put_pid(PROC_I(inode)->pid);

	/* Let go of any associated proc directory entry */
	de = PROC_I(inode)->pde;
	if (de)
		pde_put(de);
	head = PROC_I(inode)->sysctl;
	if (head) {
		rcu_assign_pointer(PROC_I(inode)->sysctl, NULL);
		sysctl_head_put(head);
	}
	/* Release any associated namespace */
	ns_ops = PROC_I(inode)->ns_ops;
	ns = PROC_I(inode)->ns;
	if (ns_ops && ns)
		ns_ops->put(ns);
}

static struct kmem_cache * proc_inode_cachep;

static struct inode *proc_alloc_inode(struct super_block *sb)
{
	struct proc_inode *ei;
	struct inode *inode;

	ei = (struct proc_inode *)kmem_cache_alloc(proc_inode_cachep, GFP_KERNEL);
	if (!ei)
		return NULL;
	ei->pid = NULL;
	ei->fd = 0;
	ei->op.proc_get_link = NULL;
	ei->pde = NULL;
	ei->sysctl = NULL;
	ei->sysctl_entry = NULL;
	ei->ns = NULL;
	ei->ns_ops = NULL;
	inode = &ei->vfs_inode;
	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
	return inode;
}

static void proc_i_callback(struct rcu_head *head)
{
	struct inode *inode = container_of(head, struct inode, i_rcu);
	kmem_cache_free(proc_inode_cachep, PROC_I(inode));
}

static void proc_destroy_inode(struct inode *inode)
{
	call_rcu(&inode->i_rcu, proc_i_callback);
}

static void init_once(void *foo)
{
	struct proc_inode *ei = (struct proc_inode *) foo;

	inode_init_once(&ei->vfs_inode);
}

void __init proc_init_inodecache(void)
{
	proc_inode_cachep = kmem_cache_create("proc_inode_cache",
					     sizeof(struct proc_inode),
					     0, (SLAB_RECLAIM_ACCOUNT|
						SLAB_MEM_SPREAD|SLAB_PANIC),
					     init_once);
}

static int proc_show_options(struct seq_file *seq, struct dentry *root)
{
	struct super_block *sb = root->d_sb;
	struct pid_namespace *pid = sb->s_fs_info;

	if (!gid_eq(pid->pid_gid, GLOBAL_ROOT_GID))
		seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, pid->pid_gid));
	if (pid->hide_pid != 0)
		seq_printf(seq, ",hidepid=%u", pid->hide_pid);

	return 0;
}

static const struct super_operations proc_sops = {
	.alloc_inode	= proc_alloc_inode,
	.destroy_inode	= proc_destroy_inode,
	.drop_inode	= generic_delete_inode,
	.evict_inode	= proc_evict_inode,
	.statfs		= simple_statfs,
	.remount_fs	= proc_remount,
	.show_options	= proc_show_options,
};

enum {BIAS = -1U<<31};

static inline int use_pde(struct proc_dir_entry *pde)
{
	int res = 1;
	spin_lock(&pde->pde_unload_lock);
	if (unlikely(pde->pde_users < 0))
		res = 0;
	else
		pde->pde_users++;
	spin_unlock(&pde->pde_unload_lock);
	return res;
}

static void __pde_users_dec(struct proc_dir_entry *pde)
{
	if (--pde->pde_users == BIAS)
		complete(pde->pde_unload_completion);
}

static void unuse_pde(struct proc_dir_entry *pde)
{
	spin_lock(&pde->pde_unload_lock);
	__pde_users_dec(pde);
	spin_unlock(&pde->pde_unload_lock);
}

void proc_entry_rundown(struct proc_dir_entry *de)
{
	spin_lock(&de->pde_unload_lock);
	de->pde_users += BIAS;
	/* Wait until all existing callers into module are done. */
	if (de->pde_users != BIAS) {
		DECLARE_COMPLETION_ONSTACK(c);
		de->pde_unload_completion = &c;
		spin_unlock(&de->pde_unload_lock);

		wait_for_completion(de->pde_unload_completion);

		spin_lock(&de->pde_unload_lock);
	}

	while (!list_empty(&de->pde_openers)) {
		struct pde_opener *pdeo;
		struct file *file;

		pdeo = list_first_entry(&de->pde_openers, struct pde_opener, lh);
		list_del(&pdeo->lh);
		spin_unlock(&de->pde_unload_lock);
		file = pdeo->file;
		de->proc_fops->release(file_inode(file), file);
		kfree(pdeo);
		spin_lock(&de->pde_unload_lock);
	}
	spin_unlock(&de->pde_unload_lock);
}

/* ->read_proc() users - legacy crap */
static ssize_t
proc_file_read(struct file *file, char __user *buf, size_t nbytes,
	       loff_t *ppos)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	ssize_t rv = -EIO;
	if (use_pde(pde)) {
		rv = __proc_file_read(file, buf, nbytes, ppos);
		unuse_pde(pde);
	}
	return rv;
}

static loff_t
proc_file_lseek(struct file *file, loff_t offset, int orig)
{
	loff_t retval = -EINVAL;
	switch (orig) {
	case 1:
		offset += file->f_pos;
	/* fallthrough */
	case 0:
		if (offset < 0 || offset > MAX_NON_LFS)
			break;
		file->f_pos = retval = offset;
	}
	return retval;
}

const struct file_operations proc_file_operations = {
	.llseek		= proc_file_lseek,
	.read		= proc_file_read,
};

static loff_t proc_reg_llseek(struct file *file, loff_t offset, int whence)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	loff_t rv = -EINVAL;
	if (use_pde(pde)) {
		loff_t (*llseek)(struct file *, loff_t, int);
		llseek = pde->proc_fops->llseek;
		if (!llseek)
			llseek = default_llseek;
		rv = llseek(file, offset, whence);
		unuse_pde(pde);
	}
	return rv;
}

static ssize_t proc_reg_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
{
	ssize_t (*read)(struct file *, char __user *, size_t, loff_t *);
	struct proc_dir_entry *pde = PDE(file_inode(file));
	ssize_t rv = -EIO;
	if (use_pde(pde)) {
		read = pde->proc_fops->read;
		if (read)
			rv = read(file, buf, count, ppos);
		unuse_pde(pde);
	}
	return rv;
}

static ssize_t proc_reg_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
{
	ssize_t (*write)(struct file *, const char __user *, size_t, loff_t *);
	struct proc_dir_entry *pde = PDE(file_inode(file));
	ssize_t rv = -EIO;
	if (use_pde(pde)) {
		write = pde->proc_fops->write;
		if (write)
			rv = write(file, buf, count, ppos);
		unuse_pde(pde);
	}
	return rv;
}

static unsigned int proc_reg_poll(struct file *file, struct poll_table_struct *pts)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	unsigned int rv = DEFAULT_POLLMASK;
	unsigned int (*poll)(struct file *, struct poll_table_struct *);
	if (use_pde(pde)) {
		poll = pde->proc_fops->poll;
		if (poll)
			rv = poll(file, pts);
		unuse_pde(pde);
	}
	return rv;
}

static long proc_reg_unlocked_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	long rv = -ENOTTY;
	long (*ioctl)(struct file *, unsigned int, unsigned long);
	if (use_pde(pde)) {
		ioctl = pde->proc_fops->unlocked_ioctl;
		if (ioctl)
			rv = ioctl(file, cmd, arg);
		unuse_pde(pde);
	}
	return rv;
}

#ifdef CONFIG_COMPAT
static long proc_reg_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	long rv = -ENOTTY;
	long (*compat_ioctl)(struct file *, unsigned int, unsigned long);
	if (use_pde(pde)) {
		compat_ioctl = pde->proc_fops->compat_ioctl;
		if (compat_ioctl)
			rv = compat_ioctl(file, cmd, arg);
		unuse_pde(pde);
	}
	return rv;
}
#endif

static int proc_reg_mmap(struct file *file, struct vm_area_struct *vma)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	int rv = -EIO;
	int (*mmap)(struct file *, struct vm_area_struct *);
	if (use_pde(pde)) {
		mmap = pde->proc_fops->mmap;
		if (mmap)
			rv = mmap(file, vma);
		unuse_pde(pde);
	}
	return rv;
}

static int proc_reg_open(struct inode *inode, struct file *file)
{
	struct proc_dir_entry *pde = PDE(inode);
	int rv = 0;
	int (*open)(struct inode *, struct file *);
	int (*release)(struct inode *, struct file *);
	struct pde_opener *pdeo;

	/*
	 * What for, you ask? Well, we can have open, rmmod, remove_proc_entry
	 * sequence. ->release won't be called because ->proc_fops will be
	 * cleared. Depending on complexity of ->release, consequences vary.
	 *
	 * We can't wait for mercy when close will be done for real, it's
	 * deadlockable: rmmod foo </proc/foo . So, we're going to do ->release
	 * by hand in remove_proc_entry(). For this, save opener's credentials
	 * for later.
	 */
	pdeo = kmalloc(sizeof(struct pde_opener), GFP_KERNEL);
	if (!pdeo)
		return -ENOMEM;

	if (!use_pde(pde)) {
		kfree(pdeo);
		return -ENOENT;
	}
	open = pde->proc_fops->open;
	release = pde->proc_fops->release;

	if (open)
		rv = open(inode, file);

	spin_lock(&pde->pde_unload_lock);
	if (rv == 0 && release) {
		/* To know what to release. */
		pdeo->file = file;
		/* Strictly for "too late" ->release in proc_reg_release(). */
		list_add(&pdeo->lh, &pde->pde_openers);
	} else
		kfree(pdeo);
	__pde_users_dec(pde);
	spin_unlock(&pde->pde_unload_lock);
	return rv;
}

static struct pde_opener *find_pde_opener(struct proc_dir_entry *pde,
					struct file *file)
{
	struct pde_opener *pdeo;

	list_for_each_entry(pdeo, &pde->pde_openers, lh) {
		if (pdeo->file == file)
			return pdeo;
	}
	return NULL;
}

static int proc_reg_release(struct inode *inode, struct file *file)
{
	struct proc_dir_entry *pde = PDE(inode);
	int rv = 0;
	int (*release)(struct inode *, struct file *);
	struct pde_opener *pdeo;

	spin_lock(&pde->pde_unload_lock);
	pdeo = find_pde_opener(pde, file);
	if (pde->pde_users < 0) {
		/*
		 * Can't simply exit, __fput() will think that everything is OK,
		 * and move on to freeing struct file. remove_proc_entry() will
		 * find slacker in opener's list and will try to do non-trivial
		 * things with struct file. Therefore, remove opener from list.
		 *
		 * But if opener is removed from list, who will ->release it?
		 */
		if (pdeo) {
			list_del(&pdeo->lh);
			spin_unlock(&pde->pde_unload_lock);
			rv = pde->proc_fops->release(inode, file);
			kfree(pdeo);
		} else
			spin_unlock(&pde->pde_unload_lock);
		return rv;
	}
	pde->pde_users++;
	release = pde->proc_fops->release;
	if (pdeo) {
		list_del(&pdeo->lh);
		kfree(pdeo);
	}
	spin_unlock(&pde->pde_unload_lock);

	if (release)
		rv = release(inode, file);

	unuse_pde(pde);
	return rv;
}

static const struct file_operations proc_reg_file_ops = {
	.llseek		= proc_reg_llseek,
	.read		= proc_reg_read,
	.write		= proc_reg_write,
	.poll		= proc_reg_poll,
	.unlocked_ioctl	= proc_reg_unlocked_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl	= proc_reg_compat_ioctl,
#endif
	.mmap		= proc_reg_mmap,
	.open		= proc_reg_open,
	.release	= proc_reg_release,
};

#ifdef CONFIG_COMPAT
static const struct file_operations proc_reg_file_ops_no_compat = {
	.llseek		= proc_reg_llseek,
	.read		= proc_reg_read,
	.write		= proc_reg_write,
	.poll		= proc_reg_poll,
	.unlocked_ioctl	= proc_reg_unlocked_ioctl,
	.mmap		= proc_reg_mmap,
	.open		= proc_reg_open,
	.release	= proc_reg_release,
};
#endif

struct inode *proc_get_inode(struct super_block *sb, struct proc_dir_entry *de)
{
	struct inode *inode = new_inode_pseudo(sb);

	if (inode) {
		inode->i_ino = de->low_ino;
		inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
		PROC_I(inode)->pde = de;

		if (de->mode) {
			inode->i_mode = de->mode;
			inode->i_uid = de->uid;
			inode->i_gid = de->gid;
		}
		if (de->size)
			inode->i_size = de->size;
		if (de->nlink)
			set_nlink(inode, de->nlink);
		WARN_ON(!de->proc_iops);
		inode->i_op = de->proc_iops;
		if (de->proc_fops) {
			if (S_ISREG(inode->i_mode)) {
#ifdef CONFIG_COMPAT
				if (!de->proc_fops->compat_ioctl)
					inode->i_fop =
						&proc_reg_file_ops_no_compat;
				else
#endif
					inode->i_fop = &proc_reg_file_ops;
			} else {
				inode->i_fop = de->proc_fops;
			}
		}
	} else
	       pde_put(de);
	return inode;
}

int proc_fill_super(struct super_block *s)
{
	struct inode *root_inode;

	s->s_flags |= MS_NODIRATIME | MS_NOSUID | MS_NOEXEC;
	s->s_blocksize = 1024;
	s->s_blocksize_bits = 10;
	s->s_magic = PROC_SUPER_MAGIC;
	s->s_op = &proc_sops;
	s->s_time_gran = 1;
	
	pde_get(&proc_root);
	root_inode = proc_get_inode(s, &proc_root);
	if (!root_inode) {
		pr_err("proc_fill_super: get root inode failed\n");
		return -ENOMEM;
	}

	s->s_root = d_make_root(root_inode);
	if (!s->s_root) {
		pr_err("proc_fill_super: allocate dentry failed\n");
		return -ENOMEM;
	}

	return proc_setup_self(s);
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/fs/proc/inode.c
	3	*
	4	* Copyright (C) 1991, 1992 Linus Torvalds
	5	*/
	6
	7	#include <linux/time.h>
	8	#include <linux/proc_fs.h>
	9	#include <linux/kernel.h>
97412950	10	#include <linux/pid_namespace.h>
1da177e4 LT	11	#include <linux/mm.h>
	12	#include <linux/string.h>
	13	#include <linux/stat.h>
786d7e16	14	#include <linux/completion.h>
dd23aae4	15	#include <linux/poll.h>
87ebdc00	16	#include <linux/printk.h>
1da177e4 LT	17	#include <linux/file.h>
	18	#include <linux/limits.h>
	19	#include <linux/init.h>
	20	#include <linux/module.h>
9043476f	21	#include <linux/sysctl.h>
97412950	22	#include <linux/seq_file.h>
5a0e3ad6	23	#include <linux/slab.h>
97412950	24	#include <linux/mount.h>
1da177e4	25
1da177e4 LT	26	#include <asm/uaccess.h>
1da177e4 LT	27
fee781e6	28	#include "internal.h"
1da177e4	29
8267952b	30	static void proc_evict_inode(struct inode *inode)
1da177e4 LT	31	{
1da177e4 LT	32	struct proc_dir_entry *de;
dfef6dcd	33	struct ctl_table_header *head;
6b4e306a	34	const struct proc_ns_operations *ns_ops;
bf056bfa	35	void *ns;
1da177e4	36
fef26658	37	truncate_inode_pages(&inode->i_data, 0);
dbd5768f	38	clear_inode(inode);
fef26658	39
99f89551	40	/* Stop tracking associated processes */
13b41b09	41	put_pid(PROC_I(inode)->pid);
1da177e4 LT	42
	43	/* Let go of any associated proc directory entry */
	44	de = PROC_I(inode)->pde;
99b76233	45	if (de)
135d5655	46	pde_put(de);
dfef6dcd AV	47	head = PROC_I(inode)->sysctl;
	48	if (head) {
	49	rcu_assign_pointer(PROC_I(inode)->sysctl, NULL);
	50	sysctl_head_put(head);
	51	}
6b4e306a EB	52	/* Release any associated namespace */
6b4e306a EB	53	ns_ops = PROC_I(inode)->ns_ops;
bf056bfa EB	54	ns = PROC_I(inode)->ns;
	55	if (ns_ops && ns)
	56	ns_ops->put(ns);
1da177e4 LT	57	}
1da177e4 LT	58
e18b890b	59	static struct kmem_cache * proc_inode_cachep;
1da177e4 LT	60
	61	static struct inode proc_alloc_inode(struct super_block sb)
	62	{
	63	struct proc_inode *ei;
	64	struct inode *inode;
	65
e94b1766	66	ei = (struct proc_inode *)kmem_cache_alloc(proc_inode_cachep, GFP_KERNEL);
1da177e4 LT	67	if (!ei)
1da177e4 LT	68	return NULL;
13b41b09	69	ei->pid = NULL;
aed7a6c4	70	ei->fd = 0;
1da177e4 LT	71	ei->op.proc_get_link = NULL;
1da177e4 LT	72	ei->pde = NULL;
9043476f AV	73	ei->sysctl = NULL;
9043476f AV	74	ei->sysctl_entry = NULL;
6b4e306a EB	75	ei->ns = NULL;
6b4e306a EB	76	ei->ns_ops = NULL;
1da177e4 LT	77	inode = &ei->vfs_inode;
	78	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
	79	return inode;
	80	}
	81
fa0d7e3d	82	static void proc_i_callback(struct rcu_head *head)
1da177e4	83	{
fa0d7e3d	84	struct inode *inode = container_of(head, struct inode, i_rcu);
1da177e4 LT	85	kmem_cache_free(proc_inode_cachep, PROC_I(inode));
	86	}
	87
fa0d7e3d NP	88	static void proc_destroy_inode(struct inode *inode)
	89	{
	90	call_rcu(&inode->i_rcu, proc_i_callback);
	91	}
	92
51cc5068	93	static void init_once(void *foo)
1da177e4 LT	94	{
	95	struct proc_inode ei = (struct proc_inode ) foo;
	96
a35afb83	97	inode_init_once(&ei->vfs_inode);
1da177e4	98	}
20c2df83	99
5bcd7ff9	100	void __init proc_init_inodecache(void)
1da177e4 LT	101	{
	102	proc_inode_cachep = kmem_cache_create("proc_inode_cache",
	103	sizeof(struct proc_inode),
fffb60f9	104	0, (SLAB_RECLAIM_ACCOUNT\|
040b5c6f	105	SLAB_MEM_SPREAD\|SLAB_PANIC),
20c2df83	106	init_once);
1da177e4 LT	107	}
1da177e4 LT	108
97412950 VK	109	static int proc_show_options(struct seq_file seq, struct dentry root)
97412950 VK	110	{
0499680a VK	111	struct super_block *sb = root->d_sb;
	112	struct pid_namespace *pid = sb->s_fs_info;
	113
dcb0f222 EB	114	if (!gid_eq(pid->pid_gid, GLOBAL_ROOT_GID))
dcb0f222 EB	115	seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, pid->pid_gid));
0499680a VK	116	if (pid->hide_pid != 0)
	117	seq_printf(seq, ",hidepid=%u", pid->hide_pid);
	118
97412950 VK	119	return 0;
	120	}
	121
ee9b6d61	122	static const struct super_operations proc_sops = {
1da177e4 LT	123	.alloc_inode = proc_alloc_inode,
1da177e4 LT	124	.destroy_inode = proc_destroy_inode,
1da177e4	125	.drop_inode = generic_delete_inode,
8267952b	126	.evict_inode = proc_evict_inode,
1da177e4	127	.statfs = simple_statfs,
97412950 VK	128	.remount_fs = proc_remount,
97412950 VK	129	.show_options = proc_show_options,
1da177e4 LT	130	};
1da177e4 LT	131
866ad9a7 AV	132	enum {BIAS = -1U<<31};
	133
	134	static inline int use_pde(struct proc_dir_entry *pde)
	135	{
	136	int res = 1;
	137	spin_lock(&pde->pde_unload_lock);
	138	if (unlikely(pde->pde_users < 0))
	139	res = 0;
	140	else
	141	pde->pde_users++;
	142	spin_unlock(&pde->pde_unload_lock);
	143	return res;
	144	}
	145
881adb85	146	static void __pde_users_dec(struct proc_dir_entry *pde)
786d7e16	147	{
866ad9a7	148	if (--pde->pde_users == BIAS)
786d7e16	149	complete(pde->pde_unload_completion);
881adb85 AD	150	}
881adb85 AD	151
866ad9a7	152	static void unuse_pde(struct proc_dir_entry *pde)
881adb85 AD	153	{
	154	spin_lock(&pde->pde_unload_lock);
	155	__pde_users_dec(pde);
786d7e16 AD	156	spin_unlock(&pde->pde_unload_lock);
	157	}
	158
866ad9a7	159	void proc_entry_rundown(struct proc_dir_entry *de)
786d7e16	160	{
866ad9a7 AV	161	spin_lock(&de->pde_unload_lock);
	162	de->pde_users += BIAS;
	163	/* Wait until all existing callers into module are done. */
	164	if (de->pde_users != BIAS) {
	165	DECLARE_COMPLETION_ONSTACK(c);
	166	de->pde_unload_completion = &c;
	167	spin_unlock(&de->pde_unload_lock);
786d7e16	168
866ad9a7 AV	169	wait_for_completion(de->pde_unload_completion);
	170
	171	spin_lock(&de->pde_unload_lock);
786d7e16	172	}
786d7e16	173
866ad9a7 AV	174	while (!list_empty(&de->pde_openers)) {
	175	struct pde_opener *pdeo;
	176	struct file *file;
786d7e16	177
866ad9a7 AV	178	pdeo = list_first_entry(&de->pde_openers, struct pde_opener, lh);
	179	list_del(&pdeo->lh);
	180	spin_unlock(&de->pde_unload_lock);
	181	file = pdeo->file;
	182	de->proc_fops->release(file_inode(file), file);
	183	kfree(pdeo);
	184	spin_lock(&de->pde_unload_lock);
	185	}
	186	spin_unlock(&de->pde_unload_lock);
786d7e16 AD	187	}
786d7e16 AD	188
866ad9a7 AV	189	/* ->read_proc() users - legacy crap */
	190	static ssize_t
	191	proc_file_read(struct file file, char __user buf, size_t nbytes,
	192	loff_t *ppos)
786d7e16	193	{
496ad9aa	194	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16	195	ssize_t rv = -EIO;
866ad9a7 AV	196	if (use_pde(pde)) {
	197	rv = __proc_file_read(file, buf, nbytes, ppos);
	198	unuse_pde(pde);
	199	}
	200	return rv;
	201	}
786d7e16	202
866ad9a7 AV	203	static loff_t
	204	proc_file_lseek(struct file *file, loff_t offset, int orig)
	205	{
	206	loff_t retval = -EINVAL;
	207	switch (orig) {
	208	case 1:
	209	offset += file->f_pos;
	210	/* fallthrough */
	211	case 0:
	212	if (offset < 0 \|\| offset > MAX_NON_LFS)
	213	break;
	214	file->f_pos = retval = offset;
786d7e16	215	}
866ad9a7 AV	216	return retval;
866ad9a7 AV	217	}
786d7e16	218
866ad9a7 AV	219	const struct file_operations proc_file_operations = {
	220	.llseek = proc_file_lseek,
	221	.read = proc_file_read,
	222	};
786d7e16	223
866ad9a7 AV	224	static loff_t proc_reg_llseek(struct file *file, loff_t offset, int whence)
	225	{
	226	struct proc_dir_entry *pde = PDE(file_inode(file));
	227	loff_t rv = -EINVAL;
	228	if (use_pde(pde)) {
	229	loff_t (llseek)(struct file , loff_t, int);
	230	llseek = pde->proc_fops->llseek;
	231	if (!llseek)
	232	llseek = default_llseek;
	233	rv = llseek(file, offset, whence);
	234	unuse_pde(pde);
	235	}
786d7e16 AD	236	return rv;
	237	}
	238
866ad9a7	239	static ssize_t proc_reg_read(struct file file, char __user buf, size_t count, loff_t *ppos)
786d7e16	240	{
866ad9a7	241	ssize_t (read)(struct file , char __user , size_t, loff_t );
496ad9aa	242	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16	243	ssize_t rv = -EIO;
866ad9a7 AV	244	if (use_pde(pde)) {
	245	read = pde->proc_fops->read;
	246	if (read)
	247	rv = read(file, buf, count, ppos);
	248	unuse_pde(pde);
786d7e16	249	}
866ad9a7 AV	250	return rv;
866ad9a7 AV	251	}
786d7e16	252
866ad9a7 AV	253	static ssize_t proc_reg_write(struct file file, const char __user buf, size_t count, loff_t *ppos)
	254	{
	255	ssize_t (write)(struct file , const char __user , size_t, loff_t );
	256	struct proc_dir_entry *pde = PDE(file_inode(file));
	257	ssize_t rv = -EIO;
	258	if (use_pde(pde)) {
	259	write = pde->proc_fops->write;
	260	if (write)
	261	rv = write(file, buf, count, ppos);
	262	unuse_pde(pde);
	263	}
786d7e16 AD	264	return rv;
	265	}
	266
	267	static unsigned int proc_reg_poll(struct file file, struct poll_table_struct pts)
	268	{
496ad9aa	269	struct proc_dir_entry *pde = PDE(file_inode(file));
dd23aae4	270	unsigned int rv = DEFAULT_POLLMASK;
786d7e16	271	unsigned int (poll)(struct file , struct poll_table_struct *);
866ad9a7 AV	272	if (use_pde(pde)) {
	273	poll = pde->proc_fops->poll;
	274	if (poll)
	275	rv = poll(file, pts);
	276	unuse_pde(pde);
786d7e16	277	}
786d7e16 AD	278	return rv;
	279	}
	280
	281	static long proc_reg_unlocked_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
	282	{
496ad9aa	283	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16	284	long rv = -ENOTTY;
b19dd42f	285	long (ioctl)(struct file , unsigned int, unsigned long);
866ad9a7 AV	286	if (use_pde(pde)) {
	287	ioctl = pde->proc_fops->unlocked_ioctl;
	288	if (ioctl)
	289	rv = ioctl(file, cmd, arg);
	290	unuse_pde(pde);
786d7e16	291	}
786d7e16 AD	292	return rv;
	293	}
	294
	295	#ifdef CONFIG_COMPAT
	296	static long proc_reg_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
	297	{
496ad9aa	298	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16 AD	299	long rv = -ENOTTY;
786d7e16 AD	300	long (compat_ioctl)(struct file , unsigned int, unsigned long);
866ad9a7 AV	301	if (use_pde(pde)) {
	302	compat_ioctl = pde->proc_fops->compat_ioctl;
	303	if (compat_ioctl)
	304	rv = compat_ioctl(file, cmd, arg);
	305	unuse_pde(pde);
786d7e16	306	}
786d7e16 AD	307	return rv;
	308	}
	309	#endif
	310
	311	static int proc_reg_mmap(struct file file, struct vm_area_struct vma)
	312	{
496ad9aa	313	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16 AD	314	int rv = -EIO;
786d7e16 AD	315	int (mmap)(struct file , struct vm_area_struct *);
866ad9a7 AV	316	if (use_pde(pde)) {
	317	mmap = pde->proc_fops->mmap;
	318	if (mmap)
	319	rv = mmap(file, vma);
	320	unuse_pde(pde);
786d7e16	321	}
786d7e16 AD	322	return rv;
	323	}
	324
	325	static int proc_reg_open(struct inode inode, struct file file)
	326	{
	327	struct proc_dir_entry *pde = PDE(inode);
	328	int rv = 0;
	329	int (open)(struct inode , struct file *);
881adb85 AD	330	int (release)(struct inode , struct file *);
	331	struct pde_opener *pdeo;
	332
	333	/*
	334	* What for, you ask? Well, we can have open, rmmod, remove_proc_entry
	335	* sequence. ->release won't be called because ->proc_fops will be
	336	* cleared. Depending on complexity of ->release, consequences vary.
	337	*
	338	* We can't wait for mercy when close will be done for real, it's
	339	* deadlockable: rmmod foo </proc/foo . So, we're going to do ->release
	340	* by hand in remove_proc_entry(). For this, save opener's credentials
	341	* for later.
	342	*/
	343	pdeo = kmalloc(sizeof(struct pde_opener), GFP_KERNEL);
	344	if (!pdeo)
	345	return -ENOMEM;
786d7e16	346
866ad9a7	347	if (!use_pde(pde)) {
881adb85	348	kfree(pdeo);
d2857e79	349	return -ENOENT;
786d7e16	350	}
786d7e16	351	open = pde->proc_fops->open;
881adb85	352	release = pde->proc_fops->release;
786d7e16 AD	353
	354	if (open)
	355	rv = open(inode, file);
	356
881adb85 AD	357	spin_lock(&pde->pde_unload_lock);
	358	if (rv == 0 && release) {
	359	/* To know what to release. */
881adb85 AD	360	pdeo->file = file;
881adb85 AD	361	/* Strictly for "too late" ->release in proc_reg_release(). */
881adb85 AD	362	list_add(&pdeo->lh, &pde->pde_openers);
	363	} else
	364	kfree(pdeo);
	365	__pde_users_dec(pde);
	366	spin_unlock(&pde->pde_unload_lock);
786d7e16 AD	367	return rv;
	368	}
	369
881adb85	370	static struct pde_opener find_pde_opener(struct proc_dir_entry pde,
866ad9a7	371	struct file *file)
881adb85 AD	372	{
	373	struct pde_opener *pdeo;
	374
	375	list_for_each_entry(pdeo, &pde->pde_openers, lh) {
866ad9a7	376	if (pdeo->file == file)
881adb85 AD	377	return pdeo;
	378	}
	379	return NULL;
	380	}
	381
786d7e16 AD	382	static int proc_reg_release(struct inode inode, struct file file)
	383	{
	384	struct proc_dir_entry *pde = PDE(inode);
	385	int rv = 0;
	386	int (release)(struct inode , struct file *);
881adb85	387	struct pde_opener *pdeo;
786d7e16 AD	388
786d7e16 AD	389	spin_lock(&pde->pde_unload_lock);
866ad9a7 AV	390	pdeo = find_pde_opener(pde, file);
866ad9a7 AV	391	if (pde->pde_users < 0) {
881adb85 AD	392	/*
	393	* Can't simply exit, __fput() will think that everything is OK,
	394	* and move on to freeing struct file. remove_proc_entry() will
	395	* find slacker in opener's list and will try to do non-trivial
	396	* things with struct file. Therefore, remove opener from list.
	397	*
	398	* But if opener is removed from list, who will ->release it?
	399	*/
	400	if (pdeo) {
	401	list_del(&pdeo->lh);
	402	spin_unlock(&pde->pde_unload_lock);
866ad9a7	403	rv = pde->proc_fops->release(inode, file);
881adb85 AD	404	kfree(pdeo);
	405	} else
	406	spin_unlock(&pde->pde_unload_lock);
786d7e16 AD	407	return rv;
	408	}
	409	pde->pde_users++;
	410	release = pde->proc_fops->release;
881adb85 AD	411	if (pdeo) {
	412	list_del(&pdeo->lh);
	413	kfree(pdeo);
	414	}
786d7e16 AD	415	spin_unlock(&pde->pde_unload_lock);
	416
	417	if (release)
	418	rv = release(inode, file);
	419
866ad9a7	420	unuse_pde(pde);
786d7e16 AD	421	return rv;
	422	}
	423
	424	static const struct file_operations proc_reg_file_ops = {
	425	.llseek = proc_reg_llseek,
	426	.read = proc_reg_read,
	427	.write = proc_reg_write,
	428	.poll = proc_reg_poll,
	429	.unlocked_ioctl = proc_reg_unlocked_ioctl,
	430	#ifdef CONFIG_COMPAT
	431	.compat_ioctl = proc_reg_compat_ioctl,
	432	#endif
	433	.mmap = proc_reg_mmap,
	434	.open = proc_reg_open,
	435	.release = proc_reg_release,
	436	};
	437
778f3dd5 DM	438	#ifdef CONFIG_COMPAT
	439	static const struct file_operations proc_reg_file_ops_no_compat = {
	440	.llseek = proc_reg_llseek,
	441	.read = proc_reg_read,
	442	.write = proc_reg_write,
	443	.poll = proc_reg_poll,
	444	.unlocked_ioctl = proc_reg_unlocked_ioctl,
	445	.mmap = proc_reg_mmap,
	446	.open = proc_reg_open,
	447	.release = proc_reg_release,
	448	};
	449	#endif
	450
6d1b6e4e	451	struct inode proc_get_inode(struct super_block sb, struct proc_dir_entry *de)
1da177e4	452	{
51f0885e	453	struct inode *inode = new_inode_pseudo(sb);
1da177e4	454
51f0885e LT	455	if (inode) {
51f0885e LT	456	inode->i_ino = de->low_ino;
a1d4aebb	457	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
a1d4aebb	458	PROC_I(inode)->pde = de;
5e971dce AD	459
	460	if (de->mode) {
	461	inode->i_mode = de->mode;
	462	inode->i_uid = de->uid;
	463	inode->i_gid = de->gid;
	464	}
	465	if (de->size)
	466	inode->i_size = de->size;
	467	if (de->nlink)
bfe86848	468	set_nlink(inode, de->nlink);
b6cdc731 AV	469	WARN_ON(!de->proc_iops);
b6cdc731 AV	470	inode->i_op = de->proc_iops;
5e971dce AD	471	if (de->proc_fops) {
5e971dce AD	472	if (S_ISREG(inode->i_mode)) {
778f3dd5	473	#ifdef CONFIG_COMPAT
5e971dce AD	474	if (!de->proc_fops->compat_ioctl)
	475	inode->i_fop =
	476	&proc_reg_file_ops_no_compat;
	477	else
778f3dd5	478	#endif
5e971dce AD	479	inode->i_fop = &proc_reg_file_ops;
	480	} else {
	481	inode->i_fop = de->proc_fops;
778f3dd5	482	}
786d7e16	483	}
99b76233	484	} else
135d5655	485	pde_put(de);
1da177e4	486	return inode;
d3d009cb	487	}
1da177e4	488
07543f5c	489	int proc_fill_super(struct super_block *s)
1da177e4	490	{
87e0aab3 MP	491	struct inode *root_inode;
87e0aab3 MP	492
92d03285	493	s->s_flags \|= MS_NODIRATIME \| MS_NOSUID \| MS_NOEXEC;
1da177e4 LT	494	s->s_blocksize = 1024;
	495	s->s_blocksize_bits = 10;
	496	s->s_magic = PROC_SUPER_MAGIC;
	497	s->s_op = &proc_sops;
	498	s->s_time_gran = 1;
	499
135d5655	500	pde_get(&proc_root);
87e0aab3 MP	501	root_inode = proc_get_inode(s, &proc_root);
87e0aab3 MP	502	if (!root_inode) {
87ebdc00	503	pr_err("proc_fill_super: get root inode failed\n");
87e0aab3 MP	504	return -ENOMEM;
87e0aab3 MP	505	}
1da177e4	506
87e0aab3 MP	507	s->s_root = d_make_root(root_inode);
87e0aab3 MP	508	if (!s->s_root) {
87ebdc00	509	pr_err("proc_fill_super: allocate dentry failed\n");
87e0aab3 MP	510	return -ENOMEM;
	511	}
	512
021ada7d	513	return proc_setup_self(s);
1da177e4	514	}